Bring Your Attribution Model up to the 21st Century

It’s a common, unfortunate duality in marketing that while marketing is customer-centric and segmented, back-end reporting and attribution models are not. I am often amazed that “customer-centric” marketing does not always have a complementary holistic and customer-value-focused attribution model to support it. It doesn’t have to be this way.

The measurement status quo is still last-touch attribution for most digital and cross-channel marketing efforts. Perhaps it’s politics — the channels that “win” the most budget with single-channel attribution have loud voices in keeping those old-fashioned models in place. Those that get less attention — like email and display — are often softer voices in the debate.

I recently worked with a non-profit organization looking to transform their business by truly understanding the value of the customer (donors, in this case). That knowledge would then guide their acquisition and retention efforts, and directly impact revenue (fundraising). The team knew that to scale the business, they needed much better measurement systems. The key question became, “How are different media and tactics performing and what is their contribution to conversion activity from the key target segments?”

Of course, the usual challenges of multiple database and non-integrated data complicated the approach, so the team focused first on marketing mix modeling and attribution.

More and more, I am hearing marketers — both B2B and B2C, and in this case a nonprofit — talk about a holistic attribution approach. It’s a very good thing for the industry and a way to strengthen any customer-centric approach. Both the front and back ends of marketing need to be customer-centric, which in my book means primarily that they recognize the value of the customer and treat the best customers differently.

This kind of holistic attribution model is often called “multi-touch attribution” and gives organizations flexibility in valuing interactions. By leveraging sophisticated algorithms, it provides for deeper insight across sources from weblogs, search engines, and social and mobile engagements. We know that customers interact with brands across channels. The back-end measurement tools must match that scope.

  • Single-touch attribution is usually based on the first touch or the last click in the customer journey, such as a phone call, store visit, and email message or direct mail piece. Single-touch has been the norm for marketers. It definitely has the advantage of being simple and readily available, but it ignores the richest trove of marketing data by ignoring the multi- and cross-channel interactions of today’s customer.
  • Multi-touch attribution looks at all the customer’s known touch points. While this has always been ideal, it has historically been too costly, too complex and therefore out of reach for most marketers. Advance marketing automation and campaign management technologies that have sophisticated analytics tools and services bring this into reach.

For this nonprofit organization, the starting point was a hybrid. The marketing mix attribution model started with an understanding of how major media spending was affecting donations. Early revelations found that email marketing, online display and video had a much larger impact than previously acknowledged. However, some of these media were not as effective as single channels in the key target segments as necessary to achieve business objectives.

Working with their data services platform provider and matching data back to the marketing database, they were able to score conversion files and media spend by segment. That provided the focus that guided the go-forward spending. Cookie-level tracking and targeting was employed to marry online and offline interactions, and new tracking tags allowed a view across different campaign initiatives.

All this took months to test and implement but the results were impressive and repeatable. What the team found is what we know intuitively — that there is no black and white answer and one media outlet did not perform best and one worst. Instead, spending in every channel contributed to the overall conversion in some target segments. Eliminating those channels that didn’t have “last click” impact would have also eliminated many conversions.

The attribution model reveals a matrix of activity contributing to conversions. No surprise that this worked — because the customer interactions are also a matrix and so are the media experiences themselves. When those matrices line up, then the real conversion and spend management occurred.

What are you doing to bring your attribution model up to the 21st Century marketing that is likely occurring on the front end of your marketing strategy? Please comment below.

Related reading